Finding the Body in the Brain. From Simulation Theory to Embodied Simulation. From: Goldman and His Critics, First Edition. Edited by Brian McLaughlin and Hilary K. Kornblith. © 2016

Goldman and His Critics, First Edition. Edited by Brian McLaughlin and Hilary K. Kornblith. © 2016 John Wiley & Sons, Inc. Published 2016 by John Wiley & Sons, Inc.

Finding the Body in the Brain

From Simulation Theory

to Embodied Simulationi

V i T T o r i o G a l l E S E

14

1 Introduction

I first met Alvin Goldman in the April of 1998 in Tucson, Arizona. I had been invited to give a plenary talk on mirror neurons (MNs) at the second Towards a Science of Consciousness conference. MNs are motor neurons we discovered in macaques’ motor cortex that discharge both when the action is executed and when it is observed being performed by someone else (di Pellegrino et al. 1992; Gallese et al. 1996; Rizzolatti et al. 1996). It was probably the first time that our discovery was presented to a wide multidisciplinary audience. Alvin attended the talk, asked me questions and afterwards invited me for lunch. In front of delicious Mexican food he briefly introduced me to simulation theory (ST), which back then I wasn’t acquainted with. Alvin pointed out the relevance of MNs for ST, as the former could constitute an important sub‐personal component of some form of low‐level simulation.

We decided to deepen our discussions and finally agreed upon writing a paper together, which appeared in the journal Trends in Cognitive Sciences the very same year (Gallese and Goldman 1998). In that paper we concluded that “a ‘cognitive continuity’ exists within the domain of intentional state attribution from non‐human primates to humans, and that MNs represent its neural correlate. This continuity is grounded in the ability of both human and non‐human primates to detect goals in the observed behavior of conspecifics. The capacity to understand action goals, already present in non‐human primates, relies on a process that matches the observed behavior to the action plans of the observer. […]. Action‐goal understanding […] constitutes a necessary phylogenetic stage within the evo-lutionary path leading to the fully developed mind‐reading abilities of human beings” (Gallese and Goldman 1998: 500).

0002700877.INDD 297 03/24/2016 12:37:10 PM

298 F I N d I N G T h e B o d y I N T h e B R A I N

I think such statement still to be valid: the empirical evidence on MNs accumulated during the following two decades has shown that in both macaques and humans MNs indeed do not map just movements, but also goal‐related motor acts. however, after the publication of our 1998 paper I developed the idea that the type of mindreading addressed by ST as conceived by Goldman was perhaps cognitively exceeding the functional prop-erties of MNs. I wanted to provide a simulation type account of mirroring that would not depend on introspection. At the same time, I felt the need to specify a simulation process that could also be applied to other neural phenomena not directly related to mindreading, like the way the cortical motor system maps space around the body, the activation of hand grasping‐related neurons during the observation of manipulable objects (canonical neu-rons) (see Murata et al. 1997; Raos et al. 2006), and the relationship between the activation of the cortical motor system and the understanding of action‐related language. Thus, I introduced the notion of embodied simulation (eS) (Gallese 2003a, 2003b, 2005a, 2005b).

In the present chapter I review these issues and clarify some aspects of eS. In particu-lar, I discuss the notion of simulation as reuse. I will conclude by introducing some recent developments of eS in relation to language, proposing that eS instantiates a form of paradigmatic knowledge. Before moving to eS we must first address how cognitive neu-roscience revolutionized our knowledge on the cortical motor system, by introducing the notion of motor cognition.

2 Motor Cognition

For decades the main goal of the neurophysiological investigation of the cortical motor system was uniquely focused on the study of elementary physical features of movement such as force, direction, and amplitude. however, a series of empirical results shows that the cortical motor system plays an important role in cognition. In particular, the neuro-physiological study of the ventral premotor cortex and the posterior parietal cortex of macaque monkeys demonstrated that the cortical motor system is functionally organized in terms of motor goals. Many cortical motor neurons do not discharge during the execu-tion of elementary movements, but are active before and during motor acts – movements executed to accomplish a specific motor outcome – such as grasping, tearing, holding or manipulating objects. These motor neurons map the relationship, in motor terms, between the agent and the object of the motor act. F5 neurons indeed become active only if a par-ticular type of effector‐object relation (for example, hand‐object) is executed until the relation leads to a different motor outcome (for example, to take possession of a piece of food, to throw it away, to break it, to bring it to the mouth, etc.) regardless of the effector employed, (see Rizzolatti et al. 2000), or of the movements the effector employs to grasp the object (Umiltà et al. 2008).

A further element of novelty about the cognitive properties of the cortical motor system concerns its role in perception, since we now know that many motor neurons are endowed with sensory properties. Several studies consistently showed that premotor and parietal areas contain neurons that perceptually respond to visual, auditory and soma-tosensory inputs (see Rizzolatti and Gallese 1997; Fogassi et al. 1996; Rizzolatti et al. 2000).

0002700877.INDD 298 03/24/2016 12:37:10 PM

F I N d I N G T h e B o d y I N T h e B R A I N 299

Altogether, these findings led to the formulation of the “Motor Cognition” hypothesis as a leading element for the emergence of social cognition (see Gallese et al. 2009). According to this hypothesis, cognitive abilities like the mapping of space and its per-ception, the perception of objects occupying our visual landscape, the hierarchical representation of action with respect to a distal goal, the detection of motor goals and action anticipation are possible because of the peculiar functional architecture of the motor system, organized in terms of goal‐directed motor acts. The proper develop-ment of such functional architecture likely scaffolds more cognitively sophisticated social cognitive abilities.

Whenever we look around, we are somehow aware of what is reachable and what is not. We can anticipate whether a falling object may hit us or not. We can calibrate the move-ment in space of our hand so as to be able to catch a fly. We can identify objects, and locations in space where sounds may come from with a remarkable precision. All of these perceptual qualities are not the outcome of the impression exerted by the external world on our perceptual and cognitive systems. Cognitive neuroscience tells us a different story. These perceptual qualities are the intentional correlates of the motor potentialities expressed by our situated body.

A peculiar example comes from the relationship between motor potentialities and spa-tial mapping, exemplified by macaque monkeys’ ventral premotor area F4 (Matelli et al. 1985), part of a parieto‐premotor cortical network mapping specific sensory events in the space near the body onto the neural representation of arm and head motor acts (Rizzolatti and Luppino 2001). F4 neurons not only control orienting/avoidance movements of the head and reaching movements of the upper limb, they also respond to tactile stimuli applied to the same body parts whose movements they control, and to visual and auditory stimuli, provided they occur within monkey’s peripersonal space. F4 neurons’ visual and auditory receptive fields (RFs) are body‐centered, that is, they are anchored to body parts and move along with them. Thus, perceiving a visual object or hearing a sound within peripersonal space evokes the motor simulation of the most appropriate actions towards that very same spatial location (Rizzolatti et al. 1997; Gallese 2005b).

Most interestingly, a putative human homologue of monkey area F4 was identified in the premotor cortex. Bremmer et al. (2001) demonstrated that the ventral region of human premotor cortex responds to tactile stimuli applied to the face and to visual and auditory stimuli presented within its peripersonal space. Furthermore, repetitive tran-scranial magnetic stimulation (TMS) over premotor cortex interferes with the processing of multisensory stimuli within the hand’s peripersonal space (Serino et al. 2011). These results show that the cortical motor system both in non‐human primates and humans maps the body’s motor potentialities and that such mapping enables the multisensory integration of self bodily‐related stimuli affecting the body and its surrounding space.

Another instantiation of motor cognition comes from canonical neurons. They discharge both during hand grasping of objects and their observation in absence of any detectable movement of the monkey (see Jeannerod et al.1995; Murata et al. 1997; Rizzolatti et al. 2001; Raos et al. 2006; Umiltà et al. 2007). Very often, a strict congruence has been observed between the type of grip controlled by a given neuron and the size or shape of the object effective in triggering its “visual” response. In a considerable percentage of neurons a

0002700877.INDD 299 03/24/2016 12:37:10 PM


congruence is observed between the response during the execution of a specific type of grip, and the visual response to objects that, although differing in shape, nevertheless all afford the same type of grip (see Murata et al. 1997; Raos et al. 2006). Thus, the very same neuron controlling a hand prehension suitable to grasp small objects will also fire equally well to the mere visual presentation of small objects like a small sphere, a small cone or a small cube. The objects shapes are different but they all specify a similar type of grasping.

The function of F5 canonical grasping neurons can hardly be defined in purely sensory or motor terms alone. Within the cortical motor system, objects are processed in motor relationally‐specified terms (Gallese 2000). According to the logic of such neural network, a series of physical entities, 3d objects, are identified, differentiated, and represented not in relation to their mere visual appearance, but in relation to the effect of the potential interaction with a situated potentially acting agent. This prop-erty qualifies as an intentional type of representation, although still fully within the functional architecture of the cortical motor system. The first conclusion we can draw is that canonical neurons contribute to a multimodal representation of individual‐object‐relations. The visual world is always also the horizon of our potential pragmatic relation to it (Gallese and Sinigaglia 2010).

What is remarkable here is the fact that the functionality of the motor system literally carves out a pragmatic Umwelt, dynamically surrounding our body. The profile of perip-ersonal space is not arbitrary: it maps and delimits a perceptual space expressing – and being constituted by – the motor potentialities of the body parts it surrounds. Manipulable objects, like the coffee mug sitting on my desk, are not only 3d shapes, visual forms with a given size, orientation, color, texture, and contrast. The coffee mug I am looking at now is the potential target of my intentional action and it is mapped as such by my cortical motor system. I submit that an important component of my perceptual experience of the coffee mug is determined, constrained, and ultimately constituted by the limits imposed by what my body can potentially do with it.

This evidence enables one to appreciate how the brain can map intentional actions. Such mapping appears to be more “basic” with respect to the standard propositional account of the representation of action. The intentional character, the “aboutness” of the representational format of our mind could be deeply rooted in the intrinsic relational character of the bodily format of bodily action representation. This, in turn, shows how intrinsically intertwined action, perception, and cognition are (hurley 1998; see also Gallese 2000).

Content is not exhausted by the propositional format of representation. Representational content cannot be fully explained without considering the ongoing modeling process of organisms as currently integrated with the object to be represented, by intending it. This integration process between the representing organism and the represented object is articulated in multiple fashions, for example, by intending to explore it by moving the eyes, by walking towards it, by intending to hold it in the focus of attention, by intending to grasp it, and ultimately, by thinking about it (see Gallese 2000; Gallese and Metzinger 2003).

The same motor circuits that control the ongoing behavior of individuals within their environment also map distances, locations and objects in that very same environment,

0002700877.INDD 300 03/24/2016 12:37:10 PM


thus defining and shaping in motor terms their representational content. The way the visual world is represented by the motor system incorporates agents’ idiosyncratic way to interact with it. To put it simply, the producer and repository of representational content is not the brain per se, but the brain‐body system, by means of its interactions with the world of which it is part.

I think these aspects already in themselves justify the necessity to define a functional process, neither confined to mindreading, nor committed to a propositional representa-tional format. eS is meant to be this sort of functional mechanism.

3 Mirroring Mechanisms

The empirical investigation of the neural basis of social cognition has been one of the most important targets of cognitive neuroscience during the last two decades. This research has repeatedly shown that a series of cortical regions (for example, mesial frontal areas, the temporo‐parietal junction, etc.) are activated during explicit mentalizing tasks (for review, see Frith and Frith 2010). Unfortunately, most of these studies do not go beyond a mere correlational enterprise. The truth is that we do not know why these corti-cal areas are relevant to mindreading; unless we content ourselves with the tautological statement that mindreading is implemented in those brain areas.

Two further problems of this approach are the reification of mental notions like intention, desire, and belief into things to be found at specific brain locations, and the questionable mindreading specificity underlying the activation of the same brain regions. A possible way out of this impasse may stem from a comparative perspective on social cognition, enabling the study of the neurophysiological mechanisms implicated in basic aspects of social cognition in non‐human primates like macaque monkeys (Gallese 2007; see also Gallese 2014; Ammaniti and Gallese 2014). As will be shown in this chapter, the investigation of the functional properties of the cortical motor system of macaques turned out to be quite fruitful.

The discovery in the early 1990s of MNs in the brain of macaques (Gallese et al. 1996; Rizzolatti et al. 1996), and the subsequent discovery of mirror mechanisms (MMs) in the human brain (see Gallese et al. 2004; Rizzolatti and Sinigaglia 2010) demonstrate that a direct modality of access to the meaning of others’ behavior is available, a modality that is different from the explicit attribution of propositional attitudes. MNs are motor neurons that not only respond to the execution of movements and actions, but also during their perception when executed by others. The relational character of behavior as mapped by the cortical motor system enables the appreciation of purpose without relying on explicit propositional inference.

The relation between actions and their outcomes was traditionally assumed to be largely independent of the motor processes and representations underpinning action execution. Such processes and representations would concern elementary motor features, such as joint displacements or muscle contractions. however, empirical evidence chal-lenges this view. We have seen that motor processes may involve motor representations of action outcomes (for example, to grasp, to place, etc.), and not only kinematic or dynamic

0002700877.INDD 301 03/24/2016 12:37:10 PM


components of actions. This suggests that beliefs, desires, and intentions are neither primitive, nor the only bearers of intentionality in action. We do not necessarily need to meta‐represent in propositional format the motor intentions of others to understand them. Motor outcomes and motor intentions are part of the “vocabulary” spoken by the motor system. In several situations we do not explicitly ascribe intentions to others; we simply detect them. Indeed, I posited that bodily formatted motor representation is enough to ground the directedness of an action to its outcome (Gallese 2000, 2003a, 2003b; see also Butterfill and Sinigaglia 2014).

The discovery of MNs gives us a new empirically founded notion of intersubjectivity connoted first and foremost as intercorporeality – the mutual resonance of intentionally meaningful sensorimotor behaviors. The ability to understand others as intentional agents does not exclusively depend on propositional competence, but it is in the first place dependent on the relational nature of action. According to this hypothesis, it is possible to directly understand others’ basic actions by means of the motor equivalence between what others do and what the observer can do.

Intercorporeality thus becomes the primordial source of knowledge that we have of others. The motor simulation instantiated by neurons endowed with “mirror properties” is probably the neural correlate of this human faculty, describable in functional terms as an instantiation of eS (Gallese 2003a, 2005a, 2011; Gallese and Sinigaglia 2011).

Action constitutes only one dimension of the rich baggage of experiences involved in interpersonal relations. every interpersonal relation implies the sharing of a multiplicity of states like, for instance, the experience of emotions and sensations. As originally hypothesized by Goldman and Gallese (2000), empirical research demonstrated that the very same nervous structures involved in the subjective experience of emotions and sensa-tions are also active when such emotions and sensations are recognized in others. A mul-tiplicity of “mirroring” mechanisms are present in our brain. It was proposed that these mechanisms, thanks to the “intentional attunement” they generate (Gallese 2006), allow us to recognize others as our fellows, likely making intersubjective communication and mutual implicit understanding possible. The functional architecture of eS seems to constitute a basic characteristic of our brain, making possible our rich and diversified intersubjective experiences, being at the basis of our capacity to empathize with others.

4 Embodied Simulation and Simulation Theory

MNs boosted a renewed interest in simulation theories and also suggested an embodied approach to simulation (Gallese 2003a, 2003b, 2005a, 2005b, 2014). embodied simulation (eS) aimed to account for basic social interactions by means of a neurobiologically plau-sible and theoretically unitary framework. What are the main differences between eS and ST? First of all, as shown above, eS is not confined to mindreading. ST is mainly applied as a fundamental heuristic strategy for mindreading. ST claims that understanding others’ behavior usually involves pretense. People first create in themselves pretend desires, preferences and beliefs of the sort they take others to have. These are then fed into their own decision‐making mechanism, which outputs pretend decisions used to predict

0002700877.INDD 302 03/24/2016 12:37:10 PM


others’ decisions (see Goldman 2006). Simulation can also be used to retrodict mental states, that is, to identify which mental states led another individual to perform a given action. Gallese and Goldman suggested that MNs’ discharge “serves the purpose of retrodicting the target mental states, moving backwards from the observed action” (Gallese and Goldman 1998: 497), thus representing “a primitive version, or possibly a precursor in phylogeny, of a simulation heuristic that might underlie mindreading” (Gallese and Goldman 1998: 498).

Two questions remained open to further developments and discussions: 1) What kind of simulation heuristic is involved in the MM, for there does not seem to be room here for pretense, or for belief and desire attribution? 2) What kind of mindreading is rooted in the MM? eS theory was introduced also to attempt answering both of these questions.

Two different views on the core meaning of “mental simulation” are currently being proposed: simulation as resemblance and simulation as reuse. According to the first view, a mental state or process simulates another mental state or process just in case it copies, reproduces, or resembles the second state or process and in doing so performs a function (Goldman 2006; see also Gordon 1986; heal 1986; Goldman 1989; Currie and Ravenscroft 2002). The notion of simulation as resemblance seems to fit the standard story of simula-tion type mindreading. The simulator supposedly forms pretend mental states matching, as closely as possible, initial mental states of the target, and uses her own decision‐making system to generate pretend mental states which match the target’s states as closely as pos-sible (Goldman 2006, 2009).

According to the alternative view, simulation as reuse, there is mental simulation just in case the same mental state or process that is used for one purpose is reused for another purpose (hurley 2008; Gallese 2009, 2011, 2014; Gallese and Sinigaglia 2011). The main argument of the reuse view is that, on almost any story, all simulation type mindreading requires any resemblance of the mental states or processes between the simulator and the target to arise from the reuse of the simulator’s own mental states or processes. At bottom it is mental reuse, not resemblance, that drives mindreading (hurley 2008).

Let’s look more closely at what the notion of reuse entails. For quite a few years (Gallese 2000) I have been advocating a role for exaptation (Gould and Lewontin 1979) as a key explanatory element of the phylogenesis of human social cognition. exaptation refers to the shift in the course of evolution of a given trait or mechanism, which is later on reused to serve new purposes and functions. According to this view, which I sketched in the pre-ceding sections, intentionality, the aboutness of our representations is – in the first place – an exapted property of the action models instantiated by the cortical motor system (see Gallese 2000: 34). The motor system not only houses causative properties but also content properties.

I subsequently introduced the notions of “neural exploitation” and “neural reuse” (Gallese and Lakoff 2005; Gallese 2008) to refer to the newly acquired commitment of sensorimotor neural resources to language and conceptual thought. Sensorimotor systems, originally evolved to guide our interactions with the world, once decoupled from the common final motor pathway and dynamically reconnected with other cortical areas – as, among others, the prefrontal regions of the brain, can be put into the service of newly acquired cognitive skills.

0002700877.INDD 303 03/24/2016 12:37:10 PM


This perspective is gaining growing consensus as epitomized by dehaene’s “neuronal recycling” hypothesis (2005), or by Anderson’s hypothesis on “neural reuse (2010). The “neuronal recycling” hypothesis was prompted by the discovery of a cortical visual area in the human occipito‐temporal region (the visual word form area, VWFA) specifically activated by early perceptual stages of the reading process. Such specificity is clearly reading‐dependent, as it doesn’t show up in individuals who never learned to read. In these individuals VWFA is activated by other non‐language‐related visual stimuli. Since reading and writing are very late cognitive acquisitions of our species, VWFA specificity for reading cannot be genetically predetermined, but it rather exemplifies an instantiation of reuse or “recycling.”

This notion of reuse holds that a given brain area’s neural specialization for processing a certain type of sensory stimuli can also instantiate a novel use‐dependent functional specialization for different stimuli of the same sensory modality. Such hypothesis does not make any strong evolutionary claim, as reuse is basically conceived of only at the ontoge-netic level. Novel cultural habits, like writing and reading, have the potentiality to remodel in a use‐dependent way a given regional brain function in the course of one individual’s life by amplifying the set of stimuli belonging to the same sensory domain it can process.

I applied the notion of neural reuse in relation to the MM and eS as a general principle of brain function. I applied it to social cognition in general, and to language and concep-tual thought in particular (Gallese and Lakoff 2005; Gallese 2008). According to Anderson’s (2010) more systematic view, by neural reuse different brain areas participate in different functions through their dynamical engagement with different brain circuits. Furthermore, a given cognitive function can be supported by a variety of brain circuits; the newer in evolutionary terms a cognitive function is, the wider is the brain circuit underpinning it.

In contrast to dehaene, both Anderson’s and my hypotheses on neural reuse do make strong evolutionary claims as they deal with the phylogenesis of human cognitive functions, challenging the strict adaptationism heralded by evolutionary psychology. Neural reuse not only enables the cortical motor system to process and integrate perceptual stimuli, hence instantiating novel cognitive functions, but also sheds new light on the phylogenesis and ontogenesis of the vicarious experiences characterizing human intersubjectivity.

Which core notion of mental simulation better fits the standard story of mindreading is not at issue here. eS theory does not aim to provide a general notion of mental simula-tion, nor a unitary account of the different stages involved in simulation type mindread-ing. Rather, eS theory aims to explain the MM and related phenomena.

eS theory posits that the MM counts as implementing mental simulation processes primarily because brain and cognitive resources typically used for one purpose are reused for another purpose. For instance, the activation of parieto‐premotor cortical networks, which typically serve the purpose of representing and accomplishing a single motor out-come (such as grasping something), or a hierarchy of motor outcomes (such as grasping something for bringing to the mouth or for placing), might also serve the purpose of attributing the same motor goal or motor intention to others. The same holds for emo-tions and sensations. Within the anterior insula the same voxels typically underpinning the subjective experience of disgust also activate when attributing disgust to others.

0002700877.INDD 304 03/24/2016 12:37:10 PM


This does not imply that one should deny the MM its matching role. Quite the contrary. From the very beginning the MM has been interpreted as a mechanism directly matching the visual or auditory representation of observed actions, emotions or sensa-tions with the observer’s own motor, visceromotor or somatosensory representations of the same actions, emotions, and sensations, respectively. The direct matching, however, is here primarily intrapersonal, since it pertains to the mental states or processes that an individual undergoes both when planning action or experiencing emotions and sensa-tions and when observing someone else’s actions, emotions and sensations. of course, this matching may also allow for interpersonal similarity of mental states or processes, but the latter would be strictly dependent upon the interpersonal sharing of the same neural and cognitive resources (Gallese and Sinigaglia 2011). When such sharing is limited or missing, people are not fully able or are not able at all to match the mental states or pro-cesses of others because they don’t have suitable mental states or processes to reuse.

The simulational reuse of mental states and processes instantiated by eS is constitutively embodied. Goldman and de Vignemont (2009) provided a very useful taxonomy of the different notions of embodiment. Accordingly, “embodied” means that body parts, bodily actions, or body representations play a crucial role in cognition. Note, however, that body representations might be interpreted in terms of mental representations either with a bodily content (representations of the body), or with a bodily format.

eS theory makes use of a notion of embodiment according to which mental states or processes are embodied because of their bodily format. The crucial notion here is that mental representations might differ in virtue not only of their content but also of their format. Just as a map and a series of sentences might represent the same route with a different format, so mental representations might have partly overlapping contents (for example, an action outcome) while differing from one another in their format (for exam-ple, bodily instead of propositional).

The bodily format of a mental representation constrains what such mental representa-tion can represent, because of the bodily constraints posed by the specific configuration of the human body. We have seen how this applies to space, objects, and others’ behaviors and experiences. A core claim of eS theory is that similar constraints apply both to the repre-sentations of one’s own actions, emotions or sensations involved in actually acting and experiencing and also to the corresponding representations involved in observing someone else performing a given action or experiencing a given emotion or sensation. These con-straints are similar because the representations have a common bodily format. hence, eS is the reuse of mental states and processes involving representations that have a bodily format. The nature and the range of what can be achieved with eS are constrained by the bodily format of the representations involved. It should be added that while eS as reuse is fully consistent with the B‐format theory put forward by Goldman and de Vignemont (2009), this doesn’t necessarily imply that all forms of embodied cognition should be based on reuse. Indeed, the firing of a neuron or group of neurons in a premotor area as part of a plan of action or a motor command would certainly qualify as an instantiation of embod-ied cognition according to the B‐format approach, but it wouldn’t be an instance of reuse.

The bodily format also determines how eS contributes to mindreading. The term “mindreading” is almost universally employed to refer to the human ability to understand

0002700877.INDD 305 03/24/2016 12:37:10 PM


others’ expressive behavior and the causes and reasons producing it. In spite of using this term, I don’t commit myself to the notion that understanding others just consists of literally “reading their minds.” I suspect that the term mindreading might qualify differ-ent types of epistemic approaches to the other. My proposal is to consider mindreading, as conceived of in a broad sense, as a non meta‐representational way of understanding others, basically sharing a common crucial feature: the mapping of the other onto the self, reciprocated by the mapping of the self on the other. This approach to intersubjectivity qualifies as second‐person perspective (see Gallese 2014).

Mindreading as conceived of in a narrow sense, should instead qualify the type of explicit third‐person form of understanding we refer to when others’ behaviors or mental states are opaque and ambiguous, thus requiring explanations. Unfortunately, the classic approach to mindreading is to date unable to convincingly explain why a series of brain areas like medial frontal areas and the temporo‐parietal junction systematically activate during explicit mentalizing tasks, besides claiming that mindreading happens to be located there (for a detailed discussion of this point, see Ammaniti and Gallese 2014: 3–6).

I posited that eS and the underpinning MMs by means of neural reuse can constitu-tively account for the representation of the motor goals of others’ actions by reusing one’s own bodily formatted motor representations, as well as of others’ emotions and sensations by reusing one’s own visceromotor and sensorimotor representations. eS can provide a unified explanatory framework for mindreading as conceived of in the broad sense speci-fied above. our bodily acting and sensing nature appears to constitute the real transcen-dental basis upon which our experience of the social world is built.

A further element of convergence with Alvin Goldman consists in how eS is able to attribute to others its contents. Basically, there are two ways in which a given mental content can be attributed to others: the first one is explicit and representational, while the second one is implicit and functional. It has been proposed that eS makes it possible to functionally attribute mental processes and contents to others (Gallese and Sinigaglia 2011). once the attribution process is spelled out in functional terms, I guess Alvin Goldman would concede that eS could constitutively support “mindreading in the broad sense,” as defined above.

Last but not least, I share with Alvin Goldman the idea that we can’t renounce the notion of representation, provided it comes in different formats, like the bodily one. If this equates to a “moderate approach to embodied cognitive science” (see Goldman 2012), then my approach is moderate too.

Understanding others is a complex enterprise. It requires the representation of others’ proximal and distal goals, others’ emotional state, the identification of the beliefs, desires, and intentions specifying the reasons promoting behavior, and the understanding of how those reasons are linked to agents and to their behavior.

5 Body and Language: Facts and Challenges

one of the key challenges for the embodied approach to human social cognition con-sists in understanding whether and how our bodily nature determines some of the key aspects identifying the uniqueness of human language. Are linguistic activities like

0002700877.INDD 306 03/24/2016 12:37:10 PM


denying, asking or doubting anchored to bodily mechanisms? The question is open and empirical research must address this challenge in the coming years.

In the meantime, at least at a purely speculative level, let us try to delineate a possible point of contact between the anthropogenic power of language and eS.

There is indeed a way to connect the common pre‐linguistic sphere to the linguistic one (Gallese 2003b, 2007, 2008; Gallese and Lakoff 2005; Glenberg and Gallese 2011). It consists in showing that language, when it refers to the body in action, brings into play the neural resources normally used to move that very same body. Seeing someone performing an action, like grabbing an object, and listening to or reading the linguistic description of that action leads to a similar motor simulation that activates some of the same regions of our cortical motor system, including those with mirror properties, normally activated when we do perform that action.

These data on the role of eS in understanding language broadly confirm the thesis according to which the bodily, sensory, and motor dimensions play a constitutive role in language production and understanding. however, it seems that the relationship between language and body does not move along a single direction. The fact is that language is unequivocally constitutive of human nature and, as such, seems to offer us wholly human modalities of experiencing our corporeity.

In this sense, neuroscientific data on the role of eS during understanding of language also point to a complementary reading with respect to the one previously proposed. on the one hand, eS might play a crucial role in understanding language. Indeed, if one reversibly interferes with this process, for instance by means of transcranial‐TMS stimu-lation, understanding of language is jeopardized. on the other hand, language allows us, and this is unique among all living species, to fix and relive specific aspects of our bodily experience. Through language we can crystallize and relive fragments of experiences that are not topical, that is to say are not my experiences now, but become a paradigm, a model, for understanding others and us.

In the following section I briefly discuss the role of eS seen as a paradigm or model in the light of the Aristotelian notion of paradeigma.ii The possibility of hypostatizing and then reliving segments of our experiences independently of the immediate physical context, or independently of specific physical stimuli, is a possibility that only the pos-session of language allows us to experience. The faculty of language is therefore, on one side, rooted in corporeity but, in turn, it changes and shapes our way of living bodily experiences.

6. Social Cognition, Body and Language: Es as a Paradigm?

The relation between body and language was to a great extent underestimated in the last century, thanks, above all, to Chomsky’s major influence. In 1966 Chomsky published a book significantly entitled Cartesian Linguistics. It is from descartes that the idea comes that language has little to do with the body. The Cartesian thesis on the relationship between language and body implies, on one side, that the body is not a substratum and material of language and, on the other, that language is exclusively the tool to express a

0002700877.INDD 307 03/24/2016 12:37:11 PM


thought that is formed independently of language itself. According to descartes and the Cartesian tradition of which Chomsky too is part, language is the tool through which we manifest an autonomous thought preceding language – a thought structured by logic but certainly not by language, whose role is circumscribed and downsized to that of being a mere label of thoughts (cf. hinzen and Sheehan 2013 for a critical discussion of the issue).

The theses informing the Cartesian idea of language can today be challenged. Language makes meaning general, releasing it from the context, that is, from the dimensions of who, what, how, where, and when. Language provides us with a unique modality of reference to the world, allowing us at the same time to transcend contingent determinations and to define them at a different level, thanks to the use of concepts like subject, object, time, space, universal, etc. Such concepts correspond to precise grammatical structures that, most likely, contributed by co‐evolutionary dynamics to the structuring of rational thought (hinzen and Sheehan 2013).

Thanks to language we can speak of mankind without referring in particular to any of the single individuals sharing the property of belonging to the human species. We can speak of a subject aside from the individual embodiments of this attribute, etc. Language provides us with general meaning, valid for everybody but, at the same time, being nobody’s meaning.

Interestingly enough, according to Giorgio Agamben (2008) what holds “for every-body and nobody” is referable to the Greek notion of paradeigma, originally explored by Aristotle. The paradeigma is a type of argumentation that moves between individual and individual according to a form of bipolar analogical knowledge. Agamben (2008: 23–4), radicalizing Aristotle’s theses, maintains that the paradigm can only be conceived of by abandoning the dichotomy between individual and universal: the rule does not exist before the single cases to which it is applied. The rule is nothing but its own exhibition in the single cases themselves, which thus it renders intelligible.

By applying the notion of paradigm to the grammatical “rules” of language, Agamben highlights a central point: the linguistic rule derives from the suspension of the concrete denotative application: “That is to say, in order to be able to serve as an example, the syntagm must be suspended from its normal function, and, nevertheless, it is precisely through this non‐operation and this suspension that it can show how the syntagm works, can allow the formulation of the rule” (2008: 26). According to Agamben, “…in the para-digm, intelligibility does not precede the phenomenon, but is, so to speak, ‘alongside’ it (parà)” (2008: 29). In other words “…in the paradigm there is not an origin or an arché: every phenomenon is the origin, every image is archaic” (2008: 33).

In Agamben’s reading, the Aristotelian paradeigma is a good model to describe the creation of linguistic rules. Starting from Agamben’s intuition and seeking to move one step further, the hypothesis that we want to explore here is that the notion of paradeigma is a good model not only for the creation of linguistic rules but also for the definition of the embodied simulation mechanism. In this connection, simulation allows us, at a senso-rimotor level, to hypostatize and reuse what holds “for everybody and nobody.”

To understand to what extent the analogy between eS and paradeigma is plausible it is necessary to go back to Aristotle (2012). What is meant by paradeigma in Aristotelian thought and in what context does Aristotle make use of this notion? The paradeigma is a

0002700877.INDD 308 03/24/2016 12:37:11 PM


typical form of rhetorical reasoning, which Aristotle discusses both in Prior Analytics and in Rhetoric. Argumentation based on the paradeigma, for example, consists in the presen-tation by the orator of an exemplary case, based on a historical fact or a figment of the imagination, as in the case of fables. It is the juxtaposition of the present situation and the exemplary one that guides, or should guide, the actions of the person to whom the argu-mentation is addressed. Thus the paradeigma, among rhetorical argumentations, is that which goes from the particular to the particular, from an exemplary case to the present situation. Argumentation based on the paradeigma does not claim universality. The orator is not bound to offer an exhaustive number of cases justifying a universally valid conclu-sion. one case is sufficient, provided that it is particularly suitable, precisely exemplary, in relation to the context in which the argumentative discourse takes place.

A distinguishing feature of the paradeigma consists in always proceeding from what is “best known and first for us” (Aristotle and Ross, 1978, II.19), or from what is for us most immediate and most easily accessible, because being part of our baggage of experiences and knowledge. At a different level of analysis, this feature also characterizes eS. The condition for the simulation mechanism to be enacted is sharing a baggage of (motor) experiences and knowledge. eS is enacted starting from what for us is “first,” that is, what for us is known and easily accessible in terms of motor potentialities and experiences. Sharing a repertoire of practices, experiences, and sensations is therefore an essential condition, since only by starting from what is well known to us it is possible to identify analogies between our actions and those of others. We understand the other starting from our own bodily experience, which is what is “best known and first for us,” again using Aristotle’s words. on the basis of this knowledge we identify similar elements in our expe-riences and as well as in those of others.

eS, when manifested in the phenomenon of action, emotion or sensation mirroring, always involves an original I–Thou relationship in which the “Thou” is the term with respect to which the self is constituted. on the other hand, the “self ” is the basis on which immediate and implicit understanding of the “Thou” is possible. The analogy with the cognitive mechanism subtended by paradigmatic reasoning appears evident. Indeed, in the case of Aristotle’s paradeigma, an example, a particular case, is understood because it is close to our feeling, our experiences, our baggage of knowledge. And the process does not stop here. This form of understanding of a particular that is not I will lead me to new conclusions and to a deeper understanding of myself, of my particular case and of my situation. our experiences are therefore the measure from which we understand others and their experiences. And others’ experiences are for us a condition for a deeper under-standing of ourselves.

Thus, the eS underpinning my experience is also a paradeigma from which I can understand what I observe in others and draw inferences from it for others and for myself.

The embodied simulation mechanism, thus defined, is constitutive of the process of construction of meaning. In this connection, eS enacted while understanding language is not my present experience but the paradeigma in relation to which some of our linguistic expressions acquire a meaning that is rooted in the body. When we read or listen to the description of an action, the process of simulation taking place in us is not the enactment of the same action; we would be echopractic if we were unable to avoid imitating and

0002700877.INDD 309 03/24/2016 12:37:11 PM


reproducing all the actions that we see or whose description we listen to or read. According to the present hypothesis, instead, eS makes available to us an exemplary case, a model, in relation to which understanding of language is also enacted. If therefore it is true that the symbolic dimension opens up some possibilities for us and creates worlds for us which only linguistic creatures can enter, on the other hand it is also true that language strongly exploits mechanisms rooted in our corporeality. enactment of the simulation process in understanding language seems to suggest that the symbolic dimension and the bodily one cohabit in linguistic praxis.

Nevertheless, the nature of this relationship is still not entirely clear, nor are the con-fines clear between the bodily dimension and the typically or exclusively symbolic one. Can it be hypothesized that corporeal knowledge also plays a role in understanding logical operators like, for instance, negation or disjunction, or that it plays a role in understand-ing the interrogative form? The whole symbolic nature of these linguistic structures appears in some respects beyond question. Research on these issues is now open and today many wonder about the possibility of identifying mechanisms that can anchor such structures to our bodily experience. This is the real challenge for the embodied cognition approach to the role played by language in human social cognition.

Let us once more return to the Aristotelian notion of paradeigma and appraise other possible hints for substantiating the analogy with the embodied simulation mechanism. The understanding that the rhetor calls for through reasoning based on the paradeigma should lead the citizen to choose what is best for him in various circumstances. The goal of such reasoning is to determine understanding of a present situation, by analogy with a historical example or a fable, and, on the basis of this more informed knowledge, to guide man’s choices. In other words, that of the rhetorical example/paradeigma is knowledge whose main goal is practical and not theoretical.

A practical aim also characterizes embodied simulation. embodied simulation is always aimed at “navigating” in the world and, therefore, eventually at acting. It was hypothe-sized that embodied simulation allows us a direct, experiential form of understanding of other people’s actions and experiences and, on the basis of this understanding, it allows us to regulate our actions and our experiences. These goals are always practical. In some respects, the process of embodied simulation that is enacted, for instance, when reading a novel (see Wojciehowski and Gallese 2011), also has a practical aim. Literature recreates a world of emotions and experiences, the emotions and the experiences of the literary char-acters inhabiting the fictional world of the novel. The simulation mechanism helps us to “navigate” in that world, even if it is a fictitious world; it allows us to understand and, partly, to relive the emotions of the protagonists and their vicissitudes. The aim in this case is practical insofar as the simulation mechanism allows us to approach the other with a second‐person epistemic perspective (Gallese 2014).

eS makes implicit knowledge about others immediately available, with the aim of regulating our interactions with them. For example, our understanding of literary char-acters is almost always second‐person, based on the possibility of perceiving analogies between our own experiences and others’ and made possible by the hypostatization of our experiences that is achieved through the simulation mechanism (Wojciehowski and Gallese, 2011).

0002700877.INDD 310 03/24/2016 12:37:11 PM


In conclusion, what is eS if not suspension of the “concrete” application of a process? Let us think of when MNs are activated in observing actions performed by others; or of when canonical neurons are activated while we are looking at the keyboard of a computer thinking about what we want to write; or when cortical motor neurons are activated when we imagine ourselves writing on that keyboard. These responses of motor neurons manifest the activa-tion of implicit knowledge, bodily motor knowledge expressing the motor potentialities of the bodily self mapped by the motor system in terms of their motor outcomes.

Reuse of motor knowledge, in the absence of the movement that realizes it as exempli-fied by eS is an example of “paradigmatic knowledge.” Thus, eS is a case of implicit paradigmatic knowledge. According to the present hypothesis eS allows us to naturalize the notion of paradigm, anchoring it at a level of sub‐personal description, whose neural correlates we can study.

our openness to the world is constituted and made possible by a motor system predis-posing and allowing us to adapt our daily and contingent pragmatic relationships with the world against the background of a prefigured but highly flexible plan of motor intention-ality. Such a plan provides its coordination to any single contingent modality of relation with the world, in which it continues to actualize itself. This aspect is important because it shows that functional processes not specific to humans, like eS, scaffold specific aspects of human social cognition.

7 Conclusions

In this chapter I addressed and discussed the notion of eS, trying to show that a new understanding of intersubjectivity can benefit from a bottom‐up study and characteriza-tion of the nonpropositional and non meta‐representational aspects of social cognition (see Gallese 2003a, 2007).

I also proposed that eS seems to be able to naturalize the notion of paradigm, natural-izing one of the processes making language reflexivity possible, and thus contributing to “create” the human. Being a subject entails being a body that learns to express itself and to express its world thanks to the paradigm – eS – that allows one to go beyond the body while remaining anchored to it.

one key issue of the new approach to intersubjectivity proposed here is the investiga-tion of the neural bases of our capacity to be attuned to the intentional relations of others. At a basic level, our interpersonal interactions do not make explicit use of propositional attitudes. This basic level consists of eS enabling the constitution of a shared meaningful interpersonal space. The shared intersubjective space in which we live from birth consti-tutes a substantial part of our semantic space. Self and other relate to each other because are opposite extensions of the same correlative and reversible we‐centric space (Gallese 2003a). observer and observed are part of a dynamic system governed by reversible rules. By means of intentional attunement, “the other” is much more than a different represen-tational system; it becomes a bodily self, like us.

The specific use of cognitive neuroscience here proposed leads to a new take on social cognition. This new take brings about the demonstration on empirical ground of the

0002700877.INDD 311 03/24/2016 12:37:11 PM


constitutive role played in foundational aspects of social cognition by the human body, when conceived of in terms of its motor potentialities. Needless to say, this only covers a partial aspect of social cognition. however, eS also provides an epistemological model, potentially useful for the empirical investigation of the more cognitively sophisticated aspects of human social cognition. This new epistemological approach to social cognition has the merit of generating predictions about the intrinsic functional nature of our social cognitive operations, cutting across, and not being subordinated to a specific mind ontol-ogy, like that purported by the classic cognitivist approach.

I am not sure whether or how much Alvin Goldman would agree with this hypothesis. What is certain is that I would have never been able to formulate it if I hadn’t been so much influenced and inspired from Goldman’s fundamental philosophical contribution to our understanding of human social cognition.

Notes

i This work was supported by the eU Grant TeSIS and by a grant from Chiesi Foundation to Vittorio Gallese.

ii For an earlier formulation of this hypothesis, see Gallese 2013; Gallese and Cuccio 2015.

References

Agamben, G. (2008) Signatura Rerum. Sul Metodo. Bollati‐Boringhieri, Turin.Ammaniti, M.A. and Gallese, V. (2014) The birth of intersubjectivity. Psychodynamics, Neurobiology,

and The Self. Norton, New york, Ny.Anderson, M.L. (2010) Neural reuse: A fundamental reorganizing principle of the brain. Behavioral

Brain Sciences 33 (4), pp.245–66.Aristotle (2012) The Art of Rhetoric. harper Press, London.Aristotle and Ross, W.d. (1978) Aristotle’s Prior and Posterior Analytics. Clarendon Press, oxford.Bremmer, F., Schlack, A., Shah, N.J., et al. (2001) Polymodal motion processing in posterior pari-

etal and premotor cortex: A human fMRI study strongly implies equivalencies between humans and monkeys. Neuron 29 (1), pp.287–96.

Butterfill, S.A. and Sinigaglia, C. (2014) Intention and motor representation in purposive action. Philosophy and Phenomenological Research 88 (1), pp.119–45.

Chomsky, N. (1966) Cartesian Linguistics: A Chapter in the History of Rationalist Thought. University Press of America, Lanham, Md.

Currie, G. and Ravenscroft, I. (2002) Recreative Minds: Imagination in Philosophy and Psychology. oxford University Press, oxford.

dehaene, S. (2005) evolution of human cortical circuits for reading and arithmetic: the neuronal recycling hypothesis. In: dehaene, S., duhamel, J., hauser, M.d., and Rizzolatti, G. (eds) From Monkey Brain to Human Brain. MIT Press, Cambridge, MA, pp.133–57.

di Pellegrino, G., Fadiga, L., Fogassi, L., Gallese, V. and Rizzolatti, G. (1992) Understanding motor events: A neurophysiological study. Experimental Brain Research 91, 176–80.

Fogassi, L., Gallese, V., Fadiga, L., Luppino, G., Matelli, M., and Rizzolatti, G. (1996) Coding of peripersonal space in inferior premotor cortex (area F4). Journal of Neurophysiology 76 (1), pp.141–57.

0002700877.INDD 312 03/24/2016 12:37:11 PM


Frith U. and Frith C. (2010) The social brain: Allowing humans to boldly go where no other species has been. Philosophical Transactions of the Royal Society B, Biological Sciences, 365 (1537), 165–76.

Gallese, V. (2000) The inner sense of action: Agency and motor representations. Journal of Consciousness Studies 7 (10), pp.23–40.

Gallese, V. (2003a) The manifold nature of interpersonal relations: The quest for a common mecha-nism. Philosophical Transactions of the Royal Society of London B 358 (1431), pp.517–28.

Gallese, V. (2003b) A neuroscientific grasp of concepts: From control to representation. Philosophical Transactions of the Royal Society of London B 358 (1431), pp.1231–40.

Gallese, V. (2005a) embodied simulation: From neurons to phenomenal experience. Phenomenology and the Cognitive Sciences 4 (1), pp.23–48.

Gallese, V. (2005b) “Being like me”: Self‐other identity, mirror neurons and empathy. In: hurley, S. and Chater, N. (eds), Perspectives on Imitation: From Cognitive Neuroscience to Social Science (Vol.1). MIT Press, Cambridge, MA, pp.101–18.

Gallese, V. (2006) Intentional attunement: A neurophysiological perspective on social cognition and its disruption in autism. Cognitive Brain Research. 1079 (1), pp.15–24.

Gallese, V. (2007) Before and below “theory of mind”: embodied simulation and the neural corre-lates of social cognition. Philosophical Transactions of the Royal Society of London B 362 (1480), pp.659–69.

Gallese, V. (2008) Mirror neurons and the social nature of language: The neural exploitation hypoth-esis. Social Neuroscience 3 (3–4), pp.317–33.

Gallese, V. (2009) Motor abstraction: A neuroscientific account of how action goals and intentions are mapped and understood. Psychological Research 73 (4), pp.486–98.

Gallese, V. (2011) Neuroscience and Phenomenology. Phenomenology & Mind 1 (2), pp.33–48.Gallese, V. (2013) Corpo non mente. Le neuroscienze cognitive e la genesi di soggettività ed inter-

soggettività. Educazione Sentimentale 20, pp.8–24.Gallese, V. (2014) Bodily selves in relation: embodied simulation as second‐person perspective on

intersubjectivity. Philosophical Transactions of the Royal Society of London B 369 (1644).Gallese, V. and Cuccio, V. (2015) The paradigmatic body: embodied simulation, intersubjectivity and

the bodily self. In: Metzinger. T. and Windt, J.M. (eds) Open MIND. MINd Group, Frankfurt.Gallese, V. and Goldman, A. (1998) Mirror neurons and the simulation theory of mind‐reading.

Trends in Cognitive Sciences, 2, 493–501.Gallese, V. and Lakoff, G. (2005) The brain’s concepts: The role of the sensory‐motor system in

reason and language. Cognitive Neuropsychology 21 (0), pp.455–79.Gallese, V. and Metzinger, T. (2003) Motor ontology: The representational reality of goals, actions,

and selves. Philosophical Psychology 16 (3), pp.365–88.Gallese, V. and Sinigaglia, C. (2010). The bodily self as power for action. Neuropsychologia, 48 (3),

pp.746–55.Gallese, V. and Sinigaglia, C. (2011) What is so special with embodied Simulation. Trends in

Cognitive Sciences 15 (11), pp.512–19.Gallese, V., Keysers, C., and Rizzolatti, G. (2004) A unifying view of the basis of social cognition.

Trends in Cognitive Sciences 8 (9), pp.396–403.Gallese, V., Fadiga, L., Fogassi, L. and Rizzolatti G. (1996) Action recognition in the premotor

cortex. Brain 119 (2), pp.593–609.Gallese, V., Rochat, M., Cossu, G., and Sinigaglia, C. (2009) Motor cognition and its role in the

phylogeny and ontogeny of action understanding. Developmental Psychology 45 (1), pp.103–13.Glenberg, A. and Gallese, V. (2011) Action‐based language: A theory of language acquisition pro-

duction and comprehension. Cortex 48 (7), pp.905–22.Goldman, A. (1989) Interpretation psychologized. Mind and Language 4 (3), pp.161–85.Goldman, A. (2006) Simulating Minds: The Philosophy, Psychology, and Neuroscience of Mindreading.

oxford University Press, oxford.

0002700877.INDD 313 03/24/2016 12:37:11 PM


Goldman, A. (2009) Mirroring, simulating, and mindreading. Mind and Language 24 (2), pp.235–52.Goldman, A. (2012) A moderate approach to embodied cognitive science. Review of Philosophy and

Psychology 3 (1), pp.71–88.Goldman, A. and Gallese, V. (2000) Reply to Schulkin. Trends in Cognitive Sciences, 4, pp.255–6.Goldman, A. and de Vignemont, F. (2009) Is social cognition embodied? Trends Cognitive Science 13

(4), pp.154–59.Gordon, R. M. (1986) Folk psychology as simulation. Mind and Language 1 (2), pp.158–71.Gould, S. J. and Lewontin, R.C. (1979) The spandrels of San Marco and the Panglossian paradigm:

A critique of the adaptationist programme. Proceedings of the Royal Society of London 205 (1161), pp.281–8.heal, J. (1986) Replication and functionalism. In: Butterfield, J. (ed.) Language, Mind and Logic.

Cambridge University Press, Cambridge.hinzen, W. and Sheehan, M. (2013) The Philosophy of Universal Grammar. oxford University

Press, oxford.hurley, S. (1998) Consciousness in action. harvard University Press, Cambridge, MA.hurley, S. (2008) Understanding simulation. Philosophy and Phenomenological Research 77 (3),

pp.755–74.Jeannerod, M., Arbib, M.A., Rizzolatti, G. and Sakata, h. (1995) Grasping objects: The cortical

mechanisms of visuomotor transformation. Trends in Neurosciences 18 (7), 314–20.Matelli, M., Luppino, G., and Rizzolatti, G. (1985) Patterns of cytochrome oxidase activity in the

frontal agranular cortex of the macaque monkey. Behavioral Brain Research 18 (2), pp.125–137.Murata, A., Fadiga, L., Fogassi, L., Gallese, V., Raos V., and Rizzolatti, G. (1997) object representation

in the ventral premotor cortex (area F5) of the monkey. Journal of Neurophysiology 78, pp.2226–30.Raos, V., Umilta, M.A., Fogassi, L., and Gallese, V. (2006) Functional properties of grasping‐related

neurons in the ventral premotor area F5 of the macaque monkey. Journal of Neurophysiology 95 (2), pp.709–29.

Rizzolatti, G. and Gallese, V. (1997) From action to meaning. In Jean‐Luc Petit (ed.), Les Neurosciences et la Philosophie de l’Action. J. Vrin Paris, pp.217–29.

Rizzolatti, G. and Luppino, G. (2001) The cortical motor system. Neuron 31 (6), pp.889–901.Rizzolatti, G. and Sinigaglia, C. (2010) The functional role of the parieto‐frontal mirror circuit:

Interpretations and misinterpretations. Nature Review Neuroscience 11, pp.264–74.Rizzolatti, G., Fogassi, L., and Gallese, V. (2000) Cortical mechanisms subserving object grasping

and action recognition: A new view on the cortical motor functions. In: Gazzaniga, M.S. (ed.) The New Cognitive Neurosciences, 2nd edn., MIT Press, Cambridge, MA, pp.539–52.

Rizzolatti G., Fogassi, L., and Gallese V. (2001) Neurophysiological mechanisms underlying the understanding and imitation of action. Nature Neuroscience Reviews 2 (9), pp.661–70.

Rizzolatti, G., Fadiga, L., Fogassi, L., and Gallese, V. (1997) The space around us. Science 277 (5323), pp.190–1.

Rizzolatti, G., Fadiga, L., Gallese, V., and Fogassi, L. (1996) Premotor cortex and the recognition of motor actions. Cognitive Brain Research 3 (2), pp.131–41.

Serino, A., Canzoneri, e., and Avenanti, A. (2011) Fronto‐parietal areas necessary for a multisen-sory representation of peripersonal space in humans: an rTMS study. Journal Cognitive Neuroscience 23 (10), pp.2956–67.

Umiltà, M.A., Brochier, T., Spinks, R.L., and Lemon, R.N. (2007) Simultaneous recording of macaque premotor and primary motor cortex neuronal populations reveals different functional contributions to visuomotor grasp. Journal of Neurophysiology 98 (1), pp.488–501.

Umiltà, M.A., escola, L., Intskirveli, I., et al. (2008) When pliers become fingers in the monkey motor system. PNAS 105 (6), pp.2209–13.

Wojciehowski h.C. and Gallese, V. (2011) how stories make us feel. Toward an embodied narratology. California Italian Studies 2 (1), [online] http://escholarship.ucop.edu/uc/item/3jg726c2/(accessed Feb. 02, 2016).

0002700877.INDD 314 03/24/2016 12:37:11 PM

Goldman and His Critics, First Edition. Edited by Brian McLaughlin and Hilary K. Kornblith. © 2016 John Wiley & Sons, Inc. Published 2016 by John Wiley & Sons, Inc.

Vittorio Gallese recounts our meeting at a conference in 1998, where he was spreading the word about mirror neurons to a large audience that had never heard of them. Mirror neurons were discovered in Parma, Italy, by a team led by Giacomo Rizzolatti, of which Gallese is a very prominent part. When I listened to his lecture, the description of mirror neurons immediately resonated with work that I and other philosophers had done on the simulation theory of mindreading. The Parma neuroscientists had not heard of the simulation theory. But it didn’t take much for me to convince Gallese of a possible relationship, or at least homology. Together we published a paper in Trends in Cognitive Sciences that same year. It proved to be an opportune time to approach social cognition through the lens of neuroscience, surprisingly, motor neuroscience. In imme-diately succeeding years there was a great boom of activity provisionally linking mirroring to various aspects of social cognition. V.S. Ramachandran, a noted psychologist, predicted that mirror neurons would do for psychology what dNA did for biology. The paper that Gallese and I published in 1998, alongside a paper by Rizzolatti and Michael Arbib, became two of the most highly cited papers in all of psychology and neuroscience in the last decade and a half, according to Gregory hickok (2014: 23) (himself a critic of the mirror neuron literature). This was all rather astonishing for me, who entered the scene with virtually zero background in neuroscience, although I immediately learned a great deal from the Parma crowd and other scientists, as well as neuroscientifically oriented philosophers. especially helpful were Vittorio and Giacomo Rizzolatti, as well as (in France) Pierre Jacob, Frederique de Vignemont, and Marc Jeannerod. The current status of mirroring is up for debate. Skeptics abound, but their lines of criticism are often rather fuzzy. To judge by its title, The Myth of Mirror Neurons, hickok’s book implies that mirror neurons are a fiction. But the text asserts no such thing; it freely

reply to Gallese

0002700877.INDD 315 03/24/2016 12:37:11 PM

316 R e P L y T o G A L L e S e

admits their existence. his real claim is that their contribution and importance have been overrated. Probably so; but what else is new?

Given space limits, I cannot dive into the territory of mirroring and mindreading. Instead I focus on another topic Gallese introduces: embodied cognition. We haven’t collaborated on this topic, but our views have substantial convergence.

The field of embodied cognition is plagued by the fact that every contributor to it means something different by it, and usually doesn’t explain their meaning clearly. I try to be clearer. Also I dissociate myself from conceptions of embodiment that deliberately aim to offer a radical departure from classical cognitivism. Classical artificial intelligence and cognitive science follow descartes in regarding intelligence as higher‐order reason and language, phenomena quite distinct from lower‐level bodily phenomena. Theorists like Rodney Brooks urge us to view intelligence as a bottom‐up phenomenon, grounded in infra‐human systems built for coping (practically) with the environment. This echoes the perspective of Merleau‐Ponty who contended that even higher‐order intelligence is controlled by “the acting body itself,” by an “I can,” not an “I think that.” Analogously, J. J. Gibson offered an account of perception free of “representations,” in contrast with classical cognitivism, in which representations are central. My approach makes ample room for representations, and seeks no radical overthrow of classical cognitivism.

Indeed, my approach to embodiment (originally introduced in Goldman and Vignemont 2009) is not on the body per se but on its representations. The brain is full of representations of bodily matters (one’s own body specifically), or coded in body‐related‐or bodily‐derived terms. This much is not new. What is new and newsworthy is the discovery that many cognitive activities with no evident relationship with one’s own body are built upon, or derive from, (own) body‐linked representations. I make no claim that all (mental) representation is in body‐related terms, but claim that there is much more such body‐related representation than classical cognitivism acknowledges. Goldman (2012) explains the matter in terms of bodily codes or formats.

Cognition C is a specimen of embodied cognition if and only if C uses some member of a special class of codes or formats for representing and/or processing its content, viz., a body‐related code or format (B‐format).

Many codes in the mind/brain represent states of the subject’s own body (from an internal perspective). Proprioception and kinaesthesis give the brain information about one’s muscles, joints, and limb positions. Codes associated with activation of the somatosensory cortex and the motor cortex are used to represent conditions of the bodily surface and to send commands to bodily effectors (respectively). These are universally acknowledged types of body‐oriented representations. The intriguing things are new findings that reveal derivative uses of such representations to perform cognitive tasks distinct from the (original) body‐oriented tasks. They reuse bodily‐related representations to represent other matters. This is what fuels interest, in various parts of cognitive science, in the embodied cognition “movement.” here is an example drawn from the mirroring domain (by no means the only source of embodiment phenomena). Although the brain initially uses certain circuits to represent one’s own current emotion states such as fear, disgust, or

0002700877.INDD 316 03/24/2016 12:37:11 PM

anger, it also reuses those circuits to represent the same states in other people (Goldman and Sripada 2005; Goldman 2013). Such cases of derivative bodily representation greatly expand the class of what qualifies, under my definition, as “bodily cognitions.” The plausibility of this approach receives theoretical backing from approaches to neural architecture that fly under the label of “massive redeployment hypothesis” or “neural reuse,” as developed by Gallese, Michael Anderson (2010), and others.

(For further discussion of embodied cognition, see the ensuing paper in this volume by Chaz Firestone, plus my reply.)

References

Anderson, M.L. (2010) Neural reuse: A fundamental organizational principle of the brain. Behavioral and Brain Sciences 33, pp.245–66.

Goldman, A.I. (2012). A moderate approach to embodied cognitive science. Review of Philosophy and Psychology 3(1), pp.71‐88.

Goldman, A.I. (2013) Joint Ventures: Mindreading, Mirroring, and Embodied Cognition. oxford University Press, New york, Ny.

Goldman, A.I. and Sripada, C.S. (2005) Simulationist models of face‐based emotion recognition. Cognition 94, pp.193–213.

Goldman, A.I. and Vignemont, F. (2009) Is social cognition embodied? Trends in Cognitive Sciences 13(4), pp.154–9.

hickok, Gregory (2014) The Myth of Mirror Neurons. W.W. Norton, New york, Ny.

R e P L y T o G A L L e S e 317

0002700877.INDD 317 03/24/2016 12:37:11 PM

Documents

Finding the Body in the Brain. From Simulation Theory to Embodied Simulation. From: Goldman and His Critics, First Edition. Edited by Brian McLaughlin and Hilary K. Kornblith. © 2016